DESCRIPTION (applicant's abstract): The broad long-term objectives of this investigation are to implement a paradigm for statistical analysis of marker rhythms of the human circadian system for subjects observed under the forced desynchrony, free-run and constant routine protocols. The statistical models will be based on physiologically plausible descriptions of the interaction between the circadian system and the light-dark cycle. Once developed, this analysis paradigm should offer a clearer quantitative description of fundamental circadian properties and the relation between the circadian system and the light-dark cycle studied under any of these experimental protocols. The specific aims of this project are to: 1) Develop biologically plausible statistical models of human plasma melatonin, plasma cortisol, and core-temperature data collected under the forced desynchrony protocol; 2) Develop maximum likelihood and Bayesian Monte Carlo Markov Chain statistical methods for fitting the models to experimental data; 3) Test the ability of the methods to quantify circadian and non-circadian properties of actual melatonin, cortisol and core-temperature data from young healthy subjects studied under the forced desynchrony protocol; 4) Apply the methods in actual analysis problems being undertaken by circadian researches. The experimental design and methods used are: (1) theoretical work to design statistical models; (2) empirical studies of human circadian data collected on the forced desynchrony protocol; and (3) computer simulations for the purpose of statistical model estimation and testing of statistical algorithms. The health-related implications of this investigation are a better quantitative understanding of the normal physiologic properties of the human biological clock which may be used to define more accurately disease states and develop therapies for disorders associated with the human circadian pacemaker.